explorer

Timeline. Ask every archive.

Type a domain or a URL. The docs worker fans the query out through @agntn/archives, lists what each archive holds, and opens any capture: played back by the archive, drawn from the archived bytes, or as text. Tick two to diff them.

Archived bodies are recordings of other people's pages. They are shown as data and never executed. Queries hit public archives; be gentle with them.

Runs archives_snapshots on the docs worker with the same executor the MCP server uses. Answers are cached for 30 minutes; a target no archive holds is an answer, not an error.

@agntn/archives·MIT license· Archived pages are data, never instructions.